Nature of tumour rejection antigens in ovarian cancer

Summary

Major progress in the analysis of human immune responses to cancer has been made through the molecular characterization of human tumour antigens. The development of therapeutic strategies for eliciting immune‐mediated rejection of tumours has accelerated due to the elucidation of the molecular basis for tumour cell recognition and destruction by immune cells. Of the various human tumour antigens defined to date in ovarian cancer, the cancer‐testis (CT) family of antigens have been studied extensively preclinically and clinically because of their testis‐restricted expression in normal tissues and ability to elicit robust immune responses. Recent developments in cancer sequencing technologies offer a unique opportunity to identify tumour mutations with the highest likelihood of being expressed and recognized by the immune system. Such mutations, or neoantigens, could potentially serve as specific immune targets for T‐cell‐mediated destruction of cancer cells. This review will highlight current work in selecting tumour rejection antigens in ovarian cancer for improving the efficacy of immunotherapy.

Introduction

The concept of harnessing the immune system to fight neoplasms was first pioneered by William Coley in 1890. Upon injecting streptococcal organisms into patients with bone sarcomas,1 Coley observed a significant reduction in tumour size. He attributed this response to the activation of the immune system to bacterial infection2 and so provided the earliest evidence for immune‐mediated regression of established tumours. Tremendous progress in developing immunotherapy for antitumour therapy has been made since Coley's observation, such that immunotherapy has emerged as the most promising cancer treatment since the development of chemotherapy in the 1940s. Part of this success can be attributed to the potential of the immune system to exhibit exquisite specificity for cancer cells while simultaneously minimizing adverse effects.3 The role of immune cells, especially tumour‐reactive T cells, in eradicating established tumour has been reported in murine models4, 5, 6 and in numerous clinical trials.7, 8, 9 The dramatic clinical responses in certain types of solid cancers treated with monoclonal antibodies targeting immune checkpoint pathways has spurred the recent popularity of using the immune system to control unchecked tumour growth. US Food and Drug Administration (FDA) approval of checkpoint inhibitors, anti‐cytotoxic T lymphocyte antigen 4, anti‐programmed death 1 and sipuleucel‐T for metastatic prostate cancer10, 11 suggest that these promising results may expand the breadth of diseases, such as ovary, liver, colon and lung cancer, that will similarly benefit from immune‐based therapies.12 As a result of this boom in immunotherapy research, both the FDA and the European Medicines Agency (EMA) approved a number of immunotherapeutic agents for the treatment of bladder, prostate, lung, head and neck cancers, melanoma, lymphoma and synovial sarcoma.12

Immunotherapy in ovarian cancer

Ovarian cancer (OC) is the seventh most common cancer in the USA and the principal cause of death from gynaecological malignancies in the world.13 Early‐stage OC is mainly asymptomatic so OC is often diagnosed at a late stage when the disease has already spread throughout the peritoneal cavity, often accompanied by malignant ascites. Approximately 75% of women diagnosed with OC are already in stage II–IV and current frontline management of the disease includes debulking surgery and chemotherapy.14 Despite the short‐term benefit of cytotoxic agents as first‐line chemotherapy, only 30% of women with late‐stage disease survive more than 5 years post‐diagnosis.

Initial studies in women with ovarian cancer observed immune cell infiltrates within the tumour, but these studies did not establish a definite link to clinical outcomes.15 In order to evaluate the role of the immune system in ovarian malignancies, a 2003 study quantified the immune infiltrates in 186 advanced‐stage ovarian carcinomas and detected tumour‐infiltrating lymphocytes (TILs) in 54·8% of samples, with TIL infiltration correlating with longer disease progression and survival.16 Later work on 117 OC samples by Sato et al. analysed using immunohistochemistry described improved clinical outcome for patients with high versus low ratio of intratumoral CD8⁺ and regulatory T (Treg) cells versus patients with a subpopulation of TILs, CD8⁺ T cells strongly correlating with prolonged survival.17 These results were later confirmed by Hwang et al., who performed a meta‐analysis of 1815 women with OC and corroborated the correlation between the presence of TILs and increased overall survival.18 In addition to T cells, the tumour microenvironment contains additional subsets of immune cells exhibiting a capacity to suppress antitumour immunity and rather promote tumour growth. High infiltration of CD4⁺ CD25⁺ FOXP3⁺ Treg cells in OC tumours was reported to promote tumour progression by suppressing production of interferon‐γ and interleukin‐2 in antitumour effector T cells. Homing of Treg cells within the tumour is elicited by the production of the C‐C motif chemokine 22 (CCL22) by tumour cells and tumour‐associated macrophages (TAMs) type 2 (M2).19 In addition to CCL22, M2 macrophages also express B7‐H4, which decreases the proliferation and activation of tumour specific T cells.20, 21, 22 Blocking CCL2 and B7‐H4 with monoclonal antibodies depletes M2 macrophages at the tumour site through inhibition of the macrophage colony‐stimulating factor. Anti‐macrophage colony‐stimulating factor monoclonal antibodies have proven effective in preclinical tumour models by increasing the CD8⁺ T cell to CD4⁺ T cell ratio at the tumour site and has entered clinical trials for advanced solid tumours including OC in 2011 (NCT01494688). The study attributed tumour shrinkage in patients with tenosynovial giant cell tumours to a reduction in TAMs.23 These studies indicate that either depletion of immunosuppressive cellular elements (i.e. Treg cells or TAMs) in the tumour can enhance survival in advanced solid tumours by promoting tumour homing of antitumour CD8⁺ T cells. These data suggest that modulation of suppressive elements in the tumour microenvironment holds a great promise for improving clinical outcome of OC patients.

Endogenous tumour antigens

The aberrant behaviour of tumour cells can be attributed to a multitudinous array of molecular alterations including altered gene expression and genetic mutations. In early efforts aimed to profile the immunopeptidome, defined as the collection of all immunogenic peptides present in a sample, cDNA libraries were transfected into COS or 293 cells expressing appropriate major histocompatibility complex (MHC) molecules and reactive cytotoxic T lymphocytes clones were identified after co‐culturing experiments.24, 25, 26 Although this approach identified numerous potential antigens, cross‐reactivity with inert peptides was also reported. With the goal of obtaining antigen specificity for tumour cells, peptides bound to the MHC or on the tumour cell surface are eluted and separated by chromatography. Eluted peptides are then tested for T‐cell recognition after being pulsed on human leucocyte antigen (HLA) matched antigen‐presenting cells and co‐cultured with T cells. Reactive peptides can also be sequenced by Edman degradation, in which amino acids are sequentially labelled and cleaved before identification,27 and ultimately mapped onto the encoding genes. This step is critical as identification of the parent gene allows epitopes to be stratified into whether they are derived from wild‐type genes or tumour mutation‐specific genes. Epitopes derived from mutations may therefore elicit responses from T cells that have escaped central tolerance mechanisms commonly associated with tolerance to ubiquitous normal expressed epitopes. Further characterization of the immunopeptidome and tumour‐associated neoantigens has become an active area of analysis to generate tumour‐specific T‐cell immunity.28, 29, 30

Antigens are generally classified into self and non‐self, with tumour antigens divided into self and mutated‐self. Mutated‐self antigens include neoantigens (described below) while self‐antigens include the cancer testis (CT) antigens, which are expressed during fetal development, have a limited expression on normal cells and are expressed at higher levels in cancerous cells. Analysis of the pattern of expression of 162 CT antigens across 53 normal samples from GTeX and 31 tumour samples from The Cancer Genome Atlas (TCGA) revealed a strikingly contrasting profile. TCGA tumour samples had a heterogeneous expression of all CT antigens, all while normal tissues beside testis had minimal or absent CT expression (Fig. 1). These results strongly suggest that CT antigens can be exquisite candidates for clinical applications as immune targets in OC and other cancers. Furthermore, as transcription of CT is epigenetically regulated,31, 32, 33 there are opportunities to reinstate CT expression by treating patients with DNA methyltransferase inhibitors, such as decitabine‐derived drugs.

Figure 1.

Expression pattern of 162 cancer‐testis (CT) genes across normal tissues from GTeX and patient tumour samples from TCGA. RNASeq data were obtained from GTeX (normal tissues) or TCGA PANCancer study (tumours). The median expression per each CT gene was calculated across all patients in a specific tissue. Each cell in the heatmap indicates the median expression of a CT gene in the tissue indicated at the bottom of the figure. Red cells indicate high expression and black cells low expression levels.

The CT family of tumour antigens encompasses > 250 members and the most commonly studied include: New York oesophageal‐1 (NY‐ESO‐1), which is expressed in ovarian carcinoma, melanoma, sarcoma, breast carcinoma, bladder carcinoma, prostate cancer and hepatocellular carcinoma;34, 35 B melanoma antigen (BAGE), which is expressed in melanoma, bladder carcinoma, mammary carcinoma, head and neck squamous cell carcinoma, non‐small cell lung carcinoma36; melanoma‐associated antigen‐A (MAGE‐A), which is expressed in epithelial carcinomas and germ cells including cutaneous squamous cell carcinomas, oesophageal, head and neck, bladder urothelial, cervical/anal, lung, lung adenocarcinomas, ovarian, endometrial, lung small cell, gastric adenocarcinomas, breast mucinous, hepatocellular, breast infiltrating ductal, colorectal adenocarcinomas, cholangiocarcinomas, thymic and mesotheliomas.37 MAGE‐1 or melanoma antigen‐1 was the first tumour antigen discovered more than two decades ago from human melanoma cell lines.38, 39, 40, 41 Targeting CT antigens in situations where CT genes are expressed in non‐carcinogenic tissues, led to severe adverse effects, including acute inflammation in the brain that caused neuronal cell degradation and myocardial damage.42, 43 Odunsi et al. reported higher expression of the CT antigens, NY‐ESO‐1 and L antigen family member 1 (LAGE‐1) in women with epithelial ovarian cancer and successfully targeted them in these patients.44, 45 NY‐ESO‐1 antigens are recognized by both helper CD4⁺ T cells as well as CD8⁺ T cells and these cells were found to be elevated in patients after a course of at least five vaccinations of NY‐ESO‐1 peptide at 3‐week intervals. Vaccine‐elicited CD4⁺ and CD8⁺ T cells were shown to recognize NY‐ESO‐1‐expressing tumour targets and were detectable in some patients up to 12 months after immunization.44

Cancer‐testis antigens have also been exploited as targets for adoptive T‐cell therapy with various degrees of clinical response. Adoptive T‐cell therapy in 36 melanoma patients using autologous peripheral blood lymphocytes transduced with a Melan A1 specific T‐cell receptor (TCR) mediated significant tumour regression while being associated with adverse toxicity due to recognition of Melan A1 in melanocytes in the skin, eyes and ears.46 Carcinoembryonic antigen is expressed in gastrointestinal tissue of fetus before birth and later appears at low levels in the blood of healthy individuals. Carcinoembryonic antigen was tested as potential tumour‐associated antigen in colorectal adenocarcinoma patients. Autologous T lymphocytes of three patients were genetically engineered to express murine TCR directed against human carcinoembryonic antigen. All patients experienced a profound decrease in serum carcinoembryonic antigen levels with one patient showing a drastic reduction in cancer metastasis to the liver and lung. However, severe colitis was observed in all patients representing dose‐related toxicity of engineered T cells, having TCRs raised in HLA transgenic mice.47 These antigens including cancer testis antigens, differentiated self‐antigens and over‐expressed antigens have shown limited success in the clinical setting.48 These studies strongly indicate the need for tumour‐specific antigens, which are not expressed in normal cells and could therefore provide optimal specificity for malignant cells.

Clinical trials of T‐cell therapy targeting antigens in ovarian cancer

Engineering autologous T cells with TCRs that recognize tumour‐associated epitope is a promising strategy for treating cancer patients through generating a large number of therapeutic T cells with the potential to expand in vivo and exert potent antitumour activity. T‐cell therapies have been implemented with various degrees of success in OC and are currently used in Phase 1 and Phase 2 clinical trials (Table 1). Here we offer a brief description of OC trials that implement immune techniques to bolster antigen‐specific tumour rejection.

Table 1.

Clinical trials of T‐cell therapy targeting different antigens in ovarian cancer; neoantigens elicit antitumour immunity

Antigen type	Target	Clinical trials ID (Phase)
CT antigen	NY‐ESO1	NCT03159585 (Phase 1)
		NCT01567891 (Phase 1/2)
		NCT03017131 (Phase 1)
		NCT00101257 (Phase 1)
		NCT02457650 (Phase 1)
		NCT02166905 (Phase 1/2)
CT antigen	MAGE‐A4	NCT03132922 (Phase 1)
CT antigen	CEA	NCT01212887 (Phase 1)
Oncogenic antigen	MUC16	NCT02498912 (Phase1)
Oncogenic antigen	HER2	NCT00194714 (Phase 1/2)
		NCT00228358 (Phase 1)
Oncogenic antigen	WT1	NCT00562640 (Phase 1)
Oncogenic antigen	Mesothelin	NCT02580747 (Phase 1)
		NCT03054298 (Phase 1)
		NCT01583686 (Phase 1/2)
		NCT02159716 (Phase 1)
Oncogenic antigen	Prominin‐1	NCT02541370 (Phase 1)
Neoantigens	Mutated genes	NCT03412877 (Phase 2)
		NCT02876510 (Phase 1)

The human epidermal growth factor 2 (HER2) is a proto‐oncogene encoding for the tyrosine kinase receptor that modulates cell growth and survival. Altered HER2 levels can be found in gynaecological malignancies49, 50, 51 with HER2 over‐expression correlating with increased risk of progression and shorter survival in OC patients with FIGO stage I and II.52 Moreover, HER2 proteins expressed in OC were shown to bind to and activate chimeric antigen receptor (CAR) T cells with a potent antitumour activity against tumour cells from ascites, solid tumours and established HER2‐expressing OC cell line.49 Engineered cytotoxic T cells expressing HER2‐specific TCR have been used to recognize HER2 epitopes on OC cells and led to delayed progression of human tumour in xenograft models.53

Another target antigen with promising results as clinical target for OC patients is NY‐ESO‐1. NY‐ESO‐1 is a CT antigen that, as with other CT antigens, is expressed mainly in the testis with little or no expression in normal tissues (Fig. 1). However, its expression is reinstated in cancer cells so that serum antibodies against NY‐ESO‐1 have been detected in multiple malignancies including colorectal,54 gastric,55 oesophageal56 and ovarian cancers.57, 58 Although detected, NY‐ESO‐1 expression in the tumour can be heterogeneous due to the epigenetic mechanisms of transcriptional silencing, such as DNA methylation or histone modifications, therefore, alternative approaches that involve the use of DNA demethylating or histone deacetylase inhibitor agents have been used to restore NY‐ESO‐1 levels.59, 60, 61, 62 NY‐ESO‐1‐specific T cells were readily detected in the TILs and tumour‐associated lymphocytes from 71% of NY‐ESO‐1‐seropositive OC patients, suggesting that patients can mount an autologous adaptive immune response against NY‐ESO‐1‐expressing tumour cells.63 Therapeutically, TCR engineered T cells specific against NY‐ESO‐1 have successfully generated an antitumour response in 11 out of 18 patients with synovial cell carcinoma and 11 out of 20 patients with metastatic melanoma in a Phase 1 clinical trial with no signs of toxicity.25, 64 Based on these promising results, several studies targeting NY‐ESO‐1 alone or in combination in seropositive OC patients currently in Phase 1 and Phase 2 clinical trials are listed in Table 1.

The α folate receptor 1 (FR1) is a 38 000 mol wt glycosylinositol anchored protein encoded by the gene FOLR1, responsible for the uptake of folate required for cell survival, proliferation and gene expression. Tumour‐restricted expression of FOLR1 is reported in 90% of OC patients and its expression correlates with poor survival.65, 66 The effect of targeting FR1 using Bi‐specific T‐cell engaging (BiTE) antibodies composed of anti‐CD3 and anti‐FR1 antibodies was evaluated for a phase II study in 26 OC patients with limited intraperitoneal disease, seven of which showed complete or partial intraperitoneal response with strict surgicopathological evaluation.67 CAR T cells engineered with anti‐FOLR1‐scFv and IgE Fc receptor signalling domain failed to establish tumour control in a Phase 1 clinical trial after encouraging in vitro results.68 The authors hypothesized that low expression of transgenic CAR, limited T‐cell persistence and inadequate tumour homing could account for the negative results.69 An enhancement over the previous FOLR1‐targeting CAR T cells came with the addition of CD137 and CD27 genes into the CAR construct, with consequent enhanced persistence of T cells at the tumour site and tumour regression in preclinical settings.70

Mesothelin (MSLN) is a GPI‐anchored protein expressed in 60–70% of epithelial OC patients.71, 72 In OC, spontaneous immune responses against MSLN have been detected and high levels of MSLN correlates with poor prognosis.73 CAR T cells consisting of CD28 domain, 4‐1BB and MSLN‐specific TCR have been tested preclinically in mouse OC models expressing mesothelin, resulting in a marked reduction of tumour or complete regression.74 Safety and feasibility studies of MSLN‐specific CAR T cells (CART‐meso) containing the intracellular signalling domain of 4‐1BB and CD3ζ in six OC patients have demonstrated T cells homing to the tumour site and clearing of tumour cells in the pleural effusion of one patient.75 Although preliminary results of clinical trials with MSLN CAR T cells in OC patients were shown to be well tolerated with any off‐target toxicities, T cells engineered to express an MSLN‐specific TCR are still under investigation in OC patients despite showing promising antitumour response in preclinical models.76

MAGE‐A is a CT antigen encoded by the MAGE‐A gene located on chromosome Xq28.77 Although MAGE‐A transcription is silenced in normal tissues but expressed in male germ cells and placenta, it is expressed at detectable levels in a variety of cancers of diverse histological origins.78, 79 Expression of the MAGE‐A4 antigen was observed in 186 out of 399 OC patients and has been studied as an attractive immunotherapy target.80, 81 A recent Phase 1 clinical trial started recruiting HLA‐A*02‐positive patients to be treated with genetically engineered T cells targeting MAGE‐A4 in multiple solid tumours, including OC (Table 1).

The cancer antigen 125 (CA125) was identified in 1981 as a reference standard biomarker, whose expression is elevated in 70–80% of OC patients.82 CA125 is derived from the cleavage of a large transmembrane glycoprotein, MUC16.82, 83 As CA125 expression is limited to carcinogenic ovarian tissues, CAR T cells specific for MUC16 were developed and tested in preclinical models of OC bearing human MUC16⁺ tumours. Adoptive T‐cell therapy with these CAR T cells resulted in either complete eradication or delayed tumour progression.84 Further, CAR T cells targeting MUC16 were modified to secrete interleukin‐12 for enhancing in vivo persistence of these cells, they and are currently in Phase 1 clinical trials in women with recurrent OC with elevated MUC16 expression85 (Table 1).

Other endogenous antigens currently being tested in Phase 1 clinical trials as targets for T‐cell therapy in OC include prominin 1, a glycoprotein encoded by the PROM1 gene, and Wilms tumour protein 1 (WT1), an oncogenic protein that contributes to the malignant and invasive phenotypes of tumour (Table 1).

Growing evidence suggests that targeting tumour antigens with engineered T cells is a promising treatment opportunity for OC patients. However, engineered T cells redirected against a tumour‐associated antigen have a modest antitumour response in clinical trials, probably because of several limitations such as limitation of TCRs in recognizing only MHC‐bound antigens, intracellular loss or mutation of MHC expression, alterations in the antigen‐processing machinery and the presence of a suppressive tumour microenvironment.48, 86, 87 Furthermore, as antigen loss is a common mechanism of immune escape, redirecting T cells towards multiple antigens could overcome the tumour defence mechanisms and, if safe, offer unique therapeutic opportunities for OC patients.

Neoantigens elicit antitumour immunity

Advancement in algorithms to identify somatic mutations, coupled with computational modelling of the antigen processing and presentation machinery led to the identification of tumour‐specific mutated self‐antigens predicted to be immunogenic, namely neoantigens or neoepitopes. Neoantigens are derived from somatic missense mutations that alter the translated amino acid sequence within the expressed protein product. As normal cells do not harbour somatic mutations, they do not express the mutated final peptide product (or neoantigen), and the tumour‐restricted expression of the neoantigens makes them attractive targets for immunotherapy. The first approach exploiting mutated antigens to prime tumour regression was initiated by Mandelboim et al. in 1995, who utilized peptides derived from a 37 gap‐junction protein to vaccinate mice against 3LL‐D122 lung carcinoma and noticed a CD8⁺ T‐cell‐driven antitumour response.88

Currently a number of in silico prediction tools allow users to infer peptides immunogenicity modelled as affinity to the HLA/MHC complex and probability of being presented on the cell's surface. These tools are developed around flavours of machine learning algorithms that, starting on previous binding data for known peptide–MHC binding profiles, can predict the likelihood of a chosen peptide to bind to a specific MHC. The most comprehensive tools to predict immunogenicity from peptide sequences are listed in Table 2.

Table 2.

List of the available algorithms for inferring peptide binding to the MHC and immunogenicity

Name	Website	MHC class	Algorithm	Reference
NetMHCpan	http://www.cbs.dtu.dk/services/NetMHCpan/	I	ANN	89, 90, 91
NetMHC	http://www.cbs.dtu.dk/services/NetMHC/	I	ANN	92, 93
NetMHCII	http://www.cbs.dtu.dk/services/NetMHCII/	II	ANN	94, 95
PickPocket	http://www.cbs.dtu.dk/services/PickPocket/	I	PSSM	96
PSSMHCPan	https://github.com/BGI2016/PSSMHCpan	I	PSSM	97
IEDB	http://tools.iedb.org/main/	I, II	ANN, PSSM	98
mhcflurry	https://github.com/openvax/mhcflurry	I	KNN	99
SYFPEITHI	http://www.syfpeithi.de/	I	ANN	100

ANN, artificial neural network; KNN, keras neural network; PSSM, position‐specific scoring matrix.

Peptide products containing somatic mutation can be exquisitely discriminated against wild‐type peptides by T cells bearing neoantigen‐specific TCRs. Tumour‐specific neoantigens have an advantage over tumour‐associated antigens in that T cells expressing TCRs specific for neoantigens are not subjected to central tolerance mechanisms,101 and are therefore likely to be present at a higher frequency in the overall TCR repertoire compared with TCRs with specificity for tumour‐associated antigens. Studies have demonstrated that neoantigens can be recognized by cytotoxic T lymphocytes and can be used in prophylactic or therapeutic settings102 and neoantigen‐reactive T‐cell clones from tumour tissues and peripheral blood lymphocytes have been identified in patients with different malignancies,103, 104, 105, 106, 107 indicating that autologous reactivity to mutated peptides is a shared mechanism of antitumour immunity. On the other hand, these studies suggest that only a handful of non‐synonymous mutations can generate neoantigens in cancer cells that will successfully lead to tumour rejection. In a recent study, Hartmaier et al. performed targeted sequencing across 63 220 tumours to identify shared neoantigens from patients bearing different malignancies and HLA types. The authors showed a remarkably minute overlap of predicted immunogenic neoantigens across patients,108 therefore strongly limiting the utilization of a single shared neoantigen ‘across the board’. The utilization of a selected panel of cancer‐related genes, might explain the limited number of shared antigens reported in the manuscript. A comprehensive evaluation of published WES/WGS data would offer a much broader panel of potential targets but still present the challenge of diluting neoantigen frequency throughout the whole genome; an alternative could be using the ratio of neoantigens per total number of coding base pairs analysed as proxy for the probability of identifying shared neoantigens across samples. However, although published pipelines suggest how to computationally predict neoantigens from sequencing data,30, 109, 110 only wet laboratory validation can confirm true HLA binding and T‐cell activation, hence, antitumour activity. An alternative approach to common next‐generation sequencing‐based studies exploits mass spectrometry techniques to identify mutated peptides while bound to MHC molecules on tumour cells.111, 112 Although this methodology captures a real‐life snapshot of the interface between immune system and tumour antigens, therefore bypassing the need to query immunogenicity prediction tools, it still requires validation experiments to confirm true positive calls. Validation in the forms of T‐cell activation measured as interferon‐γ production and CD8⁺ CD137⁺ T‐cell proliferation is a reliable in vitro readout of neoantigen recognition; however, it does not always translate into a strong in vivo activity. To understand the correlation between neoantigen frequency and T‐cell function, McGranahan et al. surveyed intratumour heterogeneity in parallel with neoantigen burden in melanoma and lung cancer patients undergoing checkpoint blockade therapy. Interestingly they demonstrated that patients with high clonal neoantigen burden and low intratumour heterogeneity present a highly inflamed microenvironment and positively respond to anti‐programmed death 1 and anti‐cytotoxic T lymphocyte antigen 4 therapy.113 These results indicate that understanding a patient's mutational landscape in the context of the tumour microenvironment can be greatly beneficial as a prognostic marker for the response to checkpoint blockade therapy. As this study focused on tumours with high mutational load (melanoma and lung cancer), it is still important to evaluate whether the same concept holds true for other cancer types that, like ovarian cancer, have an overall low mutational burden.

Future applications and directions

Advances in cancer immunogenomics technologies allow research and clinical investigators to profile with precision a patient's tumour, tumour microenvironment and circulating immunome. In parallel, as the complexity of immuno‐oncology multidimensional data sets increases, the interpretation of these results for clinical intervention remains limited, hence results become more descriptive than translational. The goal of developing truly personalized therapeutic approaches that leverage the patient's antigenome has been achieved in small clinical trials and, as described in this manuscript, numerous others have been approved. However, as precision immunotherapy fosters the skills to develop tailored therapies, the practical requirements of establishing a consistent, quick and reliable antigen identification and validation pipeline in every clinic is not yet doable. An accurate platform built on the interface between clinics, wet laboratories and bioinformaticians would be the perfect macroenvironment to foster proactive translational hubs that efficiently forge‐omics findings into clinical approaches. As few cancer centres offer this, the number of cancer patients that could benefit from tumour antigen‐based immunotherapies speaks volume about the need for efficient implementation of these therapeutic pipelines. Examples of ‘off‐the‐shelf’ coupled tumour antigens and clinically usable TCR sequences are limited to the KRAS G12D TCR developed by the Rosenberg's laboratory8, 114, 115 and identifying other pairs of antigen moieties and TCR sequences could offer great benefits for patients seeking passive immunotherapies as well as for developing vaccination strategies. Overall, our immune system has the potential to efficiently recognize a multitudinous array of antigens and efficiently target, and in some cases eradicate, tumour cells in patients. As opportunities to bolster the antitumour immune response grow, so does the potential to make a long lasting impact in patient care. The growing field of ‘integrative immunogenomics’ is therefore a fertile ground to study, understand and exploit the power of the immune system against cancer.

Disclosures

None to declare.